Identifying low-loss plasmonic materials

Plasmonics is a research area that merges optics and nanoelectronics by confining light to the nanoscale, thereby enabling a family of novel devices.^1–6 Plasmonics is also a building block for metamaterials, which are artificial materials designed and engineered to exhibit properties beyond those found in nature.^7–11 Both plasmonics and metamaterials pave the way to many applications with unprecedented functionalities, ranging from subwavelength waveguides and optical nanoantennas to hyperlenses and light concentrators.^9–11 However, at telecommunication and visible frequencies, plasmonic devices face significant challenges because of losses related to the constituent metals that seriously limit their practical use.

Figure 1. Experimentally measured (a) real (ε′) and (b) imaginary (ε′′) parts of permittivity for indium tin oxide annealed under various conditions. Ag: Silver. N: Nitrogen. O: Oxygen.

The plasmon phenomenon typically originates from the collective oscillations of free charges in a material released by an applied electromagnetic field. Consequently, plasmonic devices generally require metallic components, which have an abundance of free electrons. These electrons provide the negative real permittivity (ε′) that is an essential property of any plasmonic material. However, metals are plagued by large losses associated with the imaginary part of permittivity (ε′′), especially in the visible and UV spectral ranges, arising primarily from interband electronic transitions. Even metals with the highest conductivities suffer from large losses at optical frequencies that are detrimental to the performance of plasmonic devices.¹² To mitigate material losses, optical gain materials can be combined with metallic structures. However, even the best gain materials available offer little compensation.¹³

One approach to this problem involves alloying two or more metals to create plasmonic materials with unique band structures that can be tuned for applications at specific frequencies. For instance, by n-doping noble metals with transition metals, absorption peaks can be shifted to a region of the spectrum unimportant for a specific application.¹⁴ This technique remains valid for low doping levels (<10% volume), but will break down as the doping levels become high enough to significantly modify the material's band structure.

Although negative permittivity is a requirement of plasmonic materials, an extremely negative value is often not desirable. Many applications call for an effective permittivity close to zero (for example, the hyperlens and optical cloak).^15–23 At the telecommunication frequency, dielectric materials rarely have positive real permittivities >10, while silver has a real permittivity of about –115. Because of this large difference in the magnitude of values between silver and dielectrics, silver must be applied in extremely thin layers to achieve an effective permittivity close to zero. But differences in surface energy make depositing continuous films of silver onto a dielectric (such as silica) below ~15nm very challenging. For this reason, we look for materials whose real permittivity is negative for the desired frequency, but not extremely large in magnitude.

Doped semiconductors have plasma frequencies that can easily be tuned by adjusting doping levels and (post) processing conditions. Semiconductors can be considered low-loss plasmonic materials if they maintain a bandgap greater than the application frequency range and have a plasma frequency close to (but below) the operating frequency. Having a bandgap greater than the energy of incident light avoids interband transition losses, while having a plasma frequency near operating frequency ensures the low and negative values of real permittivity required for many applications. Two particularly interesting plasmonic materials at the telecommunication frequency (1.5μm) that we will explore in more detail in this project are indium tin oxide and doped zinc oxides.^24,25 Applied here, these materials have very low losses and thus are promising plasmonic candidates (see Figures 1 and 2).

Figure 2. Real (a) and imaginary (b) parts of permittivity for aluminum zinc oxide (AZO) and gallium zinc oxide (GZO) obtained from parameters reported in the literature. The losses in AZO and GZO are much smaller than that of silver¹² at the telecommunication wavelength.

Although the loss in a material is a necessary indicator of performance, the real part of permittivity plays a variable role in quantifying the overall material quality for different applications. Thus, different materials will perform better in different types of devices. In addition to real and imaginary values of permittivity, other factors must be taken into account when choosing a material for a plasmonic application, including fabrication practicality and cost.

As the importance of plasmonic devices continues to grow, the need for low-loss materials will play an increasingly important role. These devices will require a variety of different materials optimized for specific applications and frequencies. We hope this paper will serve as a building block for future studies.